# Qdrant — Vector Search Engine for AI Applications

> High-performance open-source vector database for AI search and RAG. Qdrant offers advanced filtering, quantization, distributed deployment, and a rich Python/JS SDK.

## Install

Copy the content below into your project:

## Quick Use

```bash
# Docker (recommended)
docker run -p 6333:6333 qdrant/qdrant

# Or pip for embedded mode
pip install qdrant-client
```

```python
from qdrant_client import QdrantClient
from qdrant_client.models import Distance, VectorParams, PointStruct

client = QdrantClient("localhost", port=6333)

# Create collection
client.create_collection(
    collection_name="docs",
    vectors_config=VectorParams(size=1536, distance=Distance.COSINE),
)

# Insert vectors
client.upsert(
    collection_name="docs",
    points=[
        PointStruct(id=1, vector=[0.1, 0.2, ...], payload={"title": "Doc A"}),
        PointStruct(id=2, vector=[0.3, 0.4, ...], payload={"title": "Doc B"}),
    ],
)

# Search
results = client.query_points(
    collection_name="docs",
    query=[0.15, 0.25, ...],
    limit=5,
)
```

## What is Qdrant?

Qdrant (read: "quadrant") is a high-performance open-source vector database written in Rust. It is purpose-built for AI applications — semantic search, RAG pipelines, recommendation systems, and anomaly detection. Qdrant stands out with its advanced payload filtering, multiple quantization options, and production-ready distributed mode.

**Answer-Ready**: Qdrant is a Rust-based open-source vector database for AI search. Advanced payload filtering, scalar/binary/product quantization, distributed deployment, and rich SDKs. Used by major AI companies for production RAG. Qdrant Cloud for managed hosting. 22k+ GitHub stars.

**Best for**: AI teams building production search and RAG systems. **Works with**: OpenAI, Cohere, HuggingFace embeddings, LangChain, LlamaIndex. **Setup time**: Under 2 minutes.

## Core Features

### 1. Advanced Filtering

```python
from qdrant_client.models import Filter, FieldCondition, MatchValue

results = client.query_points(
    collection_name="docs",
    query=[0.15, 0.25, ...],
    query_filter=Filter(
        must=[
            FieldCondition(key="category", match=MatchValue(value="ai")),
            FieldCondition(key="year", range={"gte": 2024}),
        ]
    ),
    limit=10,
)
```

### 2. Quantization (Memory Savings)

| Method | Memory | Accuracy | Speed |
|--------|--------|----------|-------|
| None (float32) | 100% | 100% | Baseline |
| Scalar (int8) | 25% | 99%+ | Faster |
| Binary | 3% | 95%+ | Fastest |
| Product | 5-10% | 97%+ | Fast |

### 3. Sparse Vectors (Hybrid Search)

```python
from qdrant_client.models import SparseVector

# Combine dense + sparse for hybrid search
client.upsert(
    collection_name="docs",
    points=[PointStruct(
        id=1,
        vector={"dense": [0.1, 0.2, ...], "sparse": SparseVector(indices=[1, 5, 100], values=[0.5, 0.3, 0.8])},
        payload={"title": "Doc A"},
    )],
)
```

### 4. Distributed Mode

```yaml
# Multi-node cluster
storage:
  shard_number: 6
  replication_factor: 2
```

### 5. Multiple SDKs

| Language | Package |
|----------|---------|
| Python | `qdrant-client` |
| JavaScript | `@qdrant/js-client-rest` |
| Rust | `qdrant-client` |
| Go | `go.qdrant.io/client` |
| Java | `io.qdrant:client` |

## Qdrant vs Alternatives

| Feature | Qdrant | Pinecone | Weaviate | Milvus |
|---------|--------|----------|----------|--------|
| Open source | Yes | No | Yes | Yes |
| Language | Rust | - | Go | Go/C++ |
| Filtering | Advanced | Basic | GraphQL | Basic |
| Quantization | 3 methods | Limited | Limited | Yes |
| Managed cloud | Yes | Yes | Yes | Yes |

## FAQ

**Q: How does it scale?**
A: Distributed mode with sharding and replication. Handles billions of vectors across multiple nodes.

**Q: Is there a managed version?**
A: Yes, Qdrant Cloud offers managed hosting with a free tier (1GB).

**Q: Can I run it embedded (no server)?**
A: Yes, `QdrantClient(":memory:")` or `QdrantClient(path="./db")` for local file storage.

## Source & Thanks

> Created by [Qdrant](https://github.com/qdrant). Licensed under Apache 2.0.
>
> [qdrant/qdrant](https://github.com/qdrant/qdrant) — 22k+ stars

<!-- ZH -->

## 快速使用

```bash
docker run -p 6333:6333 qdrant/qdrant
```

一行启动高性能向量数据库。

## 什么是 Qdrant？

Qdrant 是 Rust 编写的开源高性能向量数据库，专为 AI 应用设计。高级过滤、多种量化、分布式部署。

**一句话总结**：Rust 开源向量数据库，高级负载过滤 + 三种量化 + 分布式 + 5 种语言 SDK，AI 搜索和 RAG 首选，22k+ stars。

**适合人群**：构建生产级搜索和 RAG 的 AI 团队。

## 核心功能

### 1. 高级过滤
支持嵌套条件、范围、匹配等复杂过滤。

### 2. 量化
Scalar/Binary/Product 三种方式，最高省 97% 内存。

### 3. 混合搜索
Dense + Sparse 向量组合。

## 常见问题

**Q: 有托管版？**
A: Qdrant Cloud，免费层 1GB。

**Q: 能嵌入式运行？**
A: 可以，内存模式或本地文件。

## 来源与致谢

> [qdrant/qdrant](https://github.com/qdrant/qdrant) — 22k+ stars, Apache 2.0

---
Source: https://tokrepo.com/en/workflows/c94ec45f-6f0e-42b9-8d75-c2dd7b5840b9
Author: AI Open Source